All mice were bred and housed in a 12-h/12-h light/dark cycle starting at 7AM. Mice were housed with food and water available ad libitum. eIF4E^S209A mice on a C57BL/6 background were gifted to us from the Sonenberg laboratory at McGill University (Furic et al., 2010), and bred at The University of Arizona or The University of Texas at Dallas to produce experimental animals. Bdnf^+/- mice were obtained from The Jackson Laboratories (strain B6.129S4-Bdnf^tm1Jae/J). All mice weighed approximately 20–25 g prior to experimental use. Genotypes of the mice were determined by polymerase chain reaction (PCR) through DNA extraction of ear clips at 3–4 weeks old. The Institutional Animal Care and Use Committees at The University of Arizona, The University of Texas at Dallas, or McGill University approved all use of animal procedures. Procedures were performed according to the guidelines provided by the International Association for the Study of Pain.

Both male and female mice were used for our behavioral studies (WT: 4 males, 2 females; eIF4E^S209A: 5 males, 3 females). Testing was performed during the hours of 9AM and 4PM. Mice were habituated in their testing chambers for approximately 1 h prior to beginning the experiment. Hindpaw mechanical thresholds were determined by using the up-down method as described in Chaplan et al. (1994) using calibrated von Frey filaments (Stoelting Company, Wood Dale, IL, United States). BDNF intrathecal injections were administered in a 5 μL volume via a 3012-gauge needle (Hylden and Wilcox, 1980). The experimenter (MNA) was blinded to the genotype of the mice.

Male mice were anesthetized with ketamine and perfused with ice-cold 1× phosphate-buffered saline (PBS) solution to flush out the blood. Tissues were then isolated and flash frozen via dry ice. Frozen tissues were placed in ice cold lysis buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1 mM EDTA pH 8.0, and 1% Triton X-100) containing protease and phosphatase inhibitors cocktails (Sigma–Aldrich) and homogenized using a pestle or sonication. Samples were centrifuged at 14,000 rpm for 15 min at 4°C and the supernatant containing protein extracts was collected. Protein concentrations were assessed using the Pierce BCA protein assay kit (ThermoFisher Scientific) as directed. A total of 10–15 μg of protein was mixed with Laemmli sample buffer (Bio-Rad) and 2-mercaptoethanol and was heated at 95°C for 5 min. Samples were loaded into each well of a 10% SDS–PAGE gel along with 15 μL of Precision plus protein kaleidoscope prestained protein standards (Bio-Rad). Proteins were transferred to a 0.45 PVDF membrane (Millipore, Billierca, MA, United States) at 30 V overnight or 85 V for 1 h at 4°C. Membranes were blocked using 5% non-fat dry milk in 1× Tris Buffer Saline-Tween (TTBS) prior to primary antibody incubation. Bands were visualized using film (Kodak) or with a Bio-Rad ChemiDoc Touch. Overexposed or saturated pixels detected by the ChemiDoc Touch were excluded from analysis. Analysis was performed using ImageJ version 1.48 or Image Lab version 6.0.

The BDNF antibodies were purchased from Developmental Studies Hybridoma Bank at the University of Iowa (mouse #9; Iowa City, IA, United States) and Sigma–Aldrich (rabbit; St. Louis, MO, United States). Phospho-eIF4E, GAPDH, and trkB antibodies were obtained from Cell Signaling Technology (Danvers, MA, United States). PAR2 agonist, 2-aminothiazol-4-yl-LIGRL-NH₂ (2at-LIGRL), was synthesized as described previously (Boitano et al., 2011). Human recombinant BDNF was purchased from R&D Systems (Minneapolis, MN, United States). Prostaglandin E₂ (PGE₂) was purchased from Cayman chemicals (Ann Arbor, MI, United States). All other chemicals were attained from ThermoFisher Scientific (Waltham, MA, United States).

Lumbar DRGs and spinal cords were isolated from 3 to 6 male mice per genotype and flash-frozen on dry ice and stored at -80°C until ready to be processed. Tissues were homogenized using a pestle and total RNA was extracted using RNAqueous Total RNA Isolation kits (ThermoFisher Scientific). RNA was subsequently treated with TURBO DNase (ThermoFisher Scientific) according to the manufacturer’s instructions. RNA concentration was measured on a NanoDrop 2000 (ThermoFisher Scientific). cDNA was synthesized using iScript Reverse Transcriptase (Bio-Rad). qRT-PCR was done using a Applied Biosystems Lightcycler 7500 real-time PCR system using iTaq Universal SYBR Green Supermix (Bio-Rad) according to the manufacturer’s instructions with three technical replicates per biological replicate (averages of the technical replicates per biological replicate are reported) using primers pairs: Gapdh forward 5′-TGACCTCAACTACATGGTCTACA-3′ and Gapdh reverse 5′-CTTCCCATTCTCGGCCTT G-3′, Bdnf cds forward 5′-GCGGCAGATAAAAAGACTGC-3′ and Bdnf cds reverse 5′-GCAGCCTTCCTTGGTGTAAC-3′, and Bdnf-201 forward 5′-TGTTGGGGAGACAAGATTTT-3′ and Bdnf-201 reverse 5′-CGTGGACGTTTACTTCTTTC-3′. Bdnf primers were the same as in Matsuoka et al. (2007). Primers were made by Integrated DNA Technologies (Coralville, IA, United States). Data were analyzed as 2^-ΔΔC_T and normalized as shown in the Section “Results.” Experiments using this method of qRT-PCR (Figures 1D,E,H, 3C–F and Supplementary Figure S3 were performed at The University of Texas at Dallas.

Lumbar and thoracic DRGs were isolated from 5 mice per genotype and flash-frozen on dry ice. The DRGs were placed in chilled lysis buffer containing: 40 mM Tris–HCl, pH 7.4, 150 mM NaCl, 5 mM MgCl₂, 100 μg/ml cycloheximide, 1 mM DTT, 8% glycerol, and RNase inhibitors (RNAsin, Promega, Madison, WI, United States), and the tissue was subjected to brief homogenization using a glass homogenizer. The homogenized material was spun at 16,000 relative centrifugal force (RCF) for 10 min at 4°C, and the supernatant was loaded on a 10–50% w/w sucrose gradient in 40 mM Tris–HCl, pH 7.4, 150 mM NaCl, 5 mM MgCl_2, 100 μg/ml cyclohexamide, and RNAsin, and centrifuged at 36,000 RPM for 2.5 h at 4°C in Optima L-80 XP ultracentrifuge (Beckman Coulter, Pasadena, CA, United States) using an SW40 rotor. Polysome analysis was performed by measuring the optical density (OD) at 254 nm using an ISCO fractionator (Teledyne ISCO, Inc., Lincoln, NE, United States). RNA was extracted from each sucrose gradient fraction using TRIzol (Life Technologies). Reverse transcription was performed using a SuperScript III Reverse-Transcriptase Kit (Life Technologies) and random hexamers (Life Technologies) according to the manufacturer’s instructions. qRT-PCRs were carried out in a LightCycler 480 system using iQ Sybr Green Supermix (Bio-Rad) according to the manufacturer’s instructions using the following primers (Bdnf-201 forward 5′-GCTTTGCGGATATTGCGAAGGGTT-3′, Bdnf-201 reverse 5′-TGGAACATTGTGGCTTTGCTGTCC-3′, ActB forward-5′ -TGTGATGGTGGGAATGGGTCAGAA-3′, ActB reverse 5′-TGTGGTGCCAGATCTTCTCCATGT-3′). Results are presented in arbitrary units as relative amounts using serial dilutions of DRG RNA as qRT-PCR concentration standards. Experiments using this method of qRT-PCR (Figure 2D) were performed at McGill University.

All data are displayed as mean ± SEM, with individual samples represented within graphs to depict the n of each group and distribution. Figures 1B,C, 3A,B represent band intensities normalized to GAPDH. Figures 1D–H, 3C–F represent gene expression of 2^-ΔΔC_T normalized to Gapdh, then to WT samples. Figure 2 displays mRNA relative amounts across sucrose gradient fractions. Figures 4A,B represents hindpaw withdrawal thresholds of each animal. GraphPad Prism 6 v 6.0 for Mac OS X was used for analysis. Statistical tests, post hoc analyses, and values for each figure are displayed in Table 1.

To test our hypothesis that Bdnf mRNA translation is regulated by eIF4E phosphorylation, we first determined the specificity of antibodies for immunodetection of BDNF protein in DRGs isolated from WT and Bdnf^+/- mice (Ernfors et al., 1994). As expected, DRGs isolated from Bdnf^+/- mice exhibited a markedly reduced expression of BDNF compared to WT DRGs (Figure 1A). As an additional control, we used liver, which expresses low levels of Bdnf mRNA (Yue et al., 2014). Protein levels of BDNF in liver lysates were low compared to DRGs (Figure 1A). The BDNF band was detected above the 20 kDa marker according to the predicted molecular weight (Supplementary Figure S1A, BDNF Ensembl, BDNF Uniprot). The same BDNF band observed in Figure 1 was used for analysis throughout this study (Supplementary Figure S1B). After testing antibody specificity, we then measured BDNF protein in lysates from lumbar DRGs taken from both WT and eIF4E^S209A mice (Figure 1B). We observed a significant decrease in BDNF protein expression in eIF4E^S209A mice compared to WT. On the other hand, the levels of the BDNF receptor, trkB, were unchanged in lysates from lumbar spinal dorsal horn between genotypes (Figure 1C). A possible explanation for this deficit in BDNF protein is decreased Bdnf mRNA transcription in eIF4E^S209A mice. To examine this possibility, we measured Bdnf mRNA levels using primers that recognize all Bdnf transcript variants by qPCR. In lumbar DRGs (Figure 1D) and lumbar spinal dorsal horn (Figure 1E), no differences were observed in Bdnf cds transcript abundance between genotypes. Additionally, there were no differences in Gapdh transcript abundance (Supplementary Figure S2). Mature Bdnf mRNA can assume a variety of different variants depending on 5′ UTR exon expression. At least 11 different 5′ UTR variants for Bdnf have been annotated and they are all encoded by different exons of varying lengths (Chiaruttini et al., 2008; Mele et al., 2015). In an effort to gain insight into why Bdnf mRNA translation is decreased in the absence of eIF4E phosphorylation, we plotted the transcript length vs. the Gibbs Free Energy (ΔG) of predicted 5′ UTR folds using the RNA structure prediction algorithm mfold¹ (Mathews et al., 1999). We found that the transcript variant that is most strongly induced by pronociceptive factors in DRG, which contains the 5′ UTR encoded by the Bdnf-201 isoform (Kim et al., 2001; Matsuoka et al., 2007; Obata et al., 2011; Salerno et al., 2012; Morioka et al., 2013; Uchida et al., 2013), is the longest and has the lowest ΔG score (Figure 1F). The predicted structure of the Bdnf-201 isoform 5′ UTR is shown in Figure 1G. We therefore measured the abundance of this transcript variant in DRG using primers specific for the 5′ UTR encoded by Bdnf-201 isoform. Again, we did not find any change in abundance of this transcript in lumbar DRGs from naïve mice of both genotypes (Figure 1H).

To directly evaluate whether eIF4E phosphorylation regulates Bdnf-201 mRNA translation efficiency in DRG neurons, we sedimented and profiled polysomes isolated from lumbar and thoracic DRGs isolated from both genotypes. To obtain sufficient sample to conduct these experiments, we pooled DRGs from 5 mice per genotype. Extracts from DRGs of eIF4E^S209A and WT mice were fractionated on sucrose density gradients (Figure 2A), and the distribution of Bdnf-201 mRNA across these gradients was determined by qRT-PCR analysis using primers recognizing Bdnf-201 transcripts. Bdnf-201 mRNA associated with lighter polysome fractions in eIF4E^S209A mice, as compared to WT mice (Figures 2B,C). Total Bdnf-201 mRNA in pooled samples was not different between the two genotypes (Figure 2D). As an additional control, ActB associated with heavy polysome fractions in both WT and eIF4E^S209A DRG samples (Figure 2E). These results indicate that Bdnf-201 mRNA translation is influenced by eIF4E phosphorylation in DRGs.

We have previously shown that protease-activated receptor 2 (PAR2)-induced hyperalgesic priming is dependent on BDNF signaling (Tillu et al., 2015) and is strongly reduced in eIF4E^S209A compared to WT mice (Moy et al., 2017). Moreover, PAR2 activation promotes increased BDNF protein and mRNA transcript abundance in DRG neurons (Bao et al., 2014). We therefore sought to determine whether PAR2-induced changes in Bdnf gene expression are altered in eIF4E^S209A mice. In WT mice, we observed that the PAR2-specific agonist 2at-LIGRL induced an increase in BDNF protein levels 24 h after hindpaw injection (Figure 3A). In contrast, PAR2 stimulation in eIF4E^S209A mice failed to induce an increase in BDNF protein in affected DRGs (Figure 3B). The relative abundance of total Bdnf transcripts was not altered in DRG following PAR2 stimulation in the hindpaw in WT (Figure 3C) or eIF4E^S209A (Figure 3D) mice. However, when we specifically measured changes in the Bdnf-201 isoform abundance, we found a large increase in both WT (Figure 3E) and eIF4E^S209A (Figure 3F) mice. No differences were observed between PAR2-induced Bdnf-201 mRNA expression between WT and eIF4E^S209A DRGs (Supplementary Figure S3). Therefore, while PAR2-induced enhancement of Bdnf-201 transcription is eIF4E phosphorylation independent, eIF4E phosphorylation is required for this change in transcription to result in enhanced Bdnf-201 mRNA translation.

Based on these results, we predicted that direct injection of BDNF into the CNS should bypass any deficit in Bdnf-201 mRNA translation in eIF4E^S209A mice and produce full expression of hyperalgesic priming. BDNF (0.1 ng) was injected intrathecally in WT and eIF4E^S209A mice and mechanical hypersensitivity was measured over the ensuing 72 h. While both genotypes displayed a significant drop in withdrawal threshold, early mechanical hypersensitivity magnitudes were decreased in eIF4E^S209A mice compared to WT in response to BDNF (Figure 1A). On the other hand, no differences between genotypes were observed at 48 and 72 h after injection (Figure 4A). When mice were later challenged with PGE₂ injection into the hindpaw, full hyperalgesic priming was clearly present in both genotypes (Figure 4B) despite the deficit in acute sensitization at early time points in eIF4E^S209A mice. This is in marked contrast to observations in eIF4E^S209A mice which show little, if any hyperalgesic priming in response to PAR2 agonist, nerve growth factor (NGF), interleukin 6 (IL-6) or carrageenan injection into the hindpaw (Moy et al., 2017). Hence, the deficit in hyperalgesic priming phenotype of eIF4E^S209A mice can be rescued by direct injection of BDNF into the spinal cord.